1. Field of the Invention
The present invention relates to a color linear image sensor and driving method, and more particularly to a color linear image sensor and its driving method for reducing residual images without greatly increasing a line-to-line distance of the color linear image sensor.
2. Description of the Related Art
With popularization of personal computers and requests for higher performance of copying machines in recent years, demands for color linear image sensors for reading color images have increased.
Such a color linear image sensor is usually constructed in a manner that three CCD linear image sensors having charge transfer functions are arranged in parallel and color filters having different colors, e.g., R (red), G (green) and B (blue) filters, are placed on the light receiving section array of each CCD linear image sensor.
FIG. 11 is an entire constitutional view showing an example of such a color linear image sensor.
In FIG. 11 reference numerals 1a, 1b and 1c denote light receiving sections having color filters of RGB (not shown) placed thereon; and 2a, 2b and 2c denote signal charge reading sections for reading signal charges stored after photoelectric conversion in the light receiving sections to signal charge transfer sections 3a, 3b and 3c provided in the vicinity. Signal charges Q1, Q2 and Q3 of the respective light receiving sections are read to the signal charge transfer sections (indicated by white arrows. Reading pulses are xcfx86TG1, 2 and 3. Pulse lines are not shown).
In the case of the CCD linear image sensor, the signal charge transfer sections 3a, 3b and 3c usually include two-phase driven CCD shift registers, and pulse lines L1, L2, L3, L4 and L5 (clock pulse is xcfx861 or xcfx862; connection from the pulse line to each signal charge transfer section is indicated by an arrow) for driving the two-phase driven CCD shift registers are arranged in the vicinity of the signal charge transfer sections.
Signal charges transferred by the signal charge transfer sections 3a, 3b and 3c are outputted to the outside by output circuits 4a, 4b and 4c composed of signal charge detection sections including a floating diffusion region for converting the signal charges into signal voltages and analog circuits such as source followers or inverters, and then color signals are obtained.
FIG. 15 is a timing chart showing a driving method of the conventional color linear image sensor shown in FIG. 11. (Common among colors).
The signal charges of the light receiving sections 1a, 1b and 1c are stored while clock pulses xcfx86TG1, 2 and 3 applied to the signal charge reading sections 2a, 2b and 2c are at LOW levels. In the period T when the clock pulses are at HIGH levels, the signal charges are read to the specified signal charge transfer sections 3a, 3b and 3c. 
Subsequently, in these signal charge transfer sections, the signal charges are transferred by two-phase clock pulses xcfx861 and xcfx862 (phases reverse to each other), and then outputted from the output circuits 4a, 4b and 4c. 
If the foregoing color linear image sensor is used for a scanner or a copying machine, scanning is performed by the three CCD linear image sensors having the color filters vertically to the arraying direction thereof. Accordingly, for obtaining color information (e.g., RGB) regarding an image of a specified place on an object, signal processing must be performed after externally storing bits of color information of first and second lines in the period from the end of scanning the specified place by the first line (e.g., R) to the end of scanning by a third line (e.g., B) and then obtaining three bits of color information. Consequently, an external memory having a very large capacity must be used.
For example, in the case of a color linear image sensor of a class of 5300 pixelsxc3x973 array which is used for a color copying machine or a high resolution color scanner, if gradation is set to 10 bits, a necessary capacity C of an external memory is obtained by the following expression:
C=5300xc3x9710xc3x973xc3x97(M+1) bitsxe2x80x83xe2x80x83(1)
Herein, M denotes, in the form of a scanning number of times, a line-to-line distance between two light receiving section array adjacent to each other among the light receiving section array. For example, if a size of one pixel of each of the R, G and B light receiving sections is 8 xcexcmxc3x978 xcexcm and line-to-line distances between the R and G light receiving section array and between the G and B light receiving section array are both 64 xcexcm, M is obtained by the following expression:
M=64 xcexcm/8 xcexcmxe2x80x83xe2x80x83(2)
Accordingly, a capacity C of the external memory is 1431000 bits.
As apparent from the expression (1), for reducing a capacity of the external memory, it is necessary to reduce a scanning number of times in the period from the end of scanning by the first line (e.g., R) to the end of scanning by the third line (e.g., B) by shortening a distance among the three light receiving section array.
FIG. 12 is an expanded view of a region surrounded with a broken line X1 of FIG. 11. In the drawing, the same reference numerals as those used in FIG. 11, e.g., 1a and 1b, denote the same elements. A reference numeral 5 denotes an aluminum wiring line to which a two-phase clock pulse xcfx861 or xcfx862 is applied; 6 a contact for connecting the aluminum wiring line 5 with one selected from two kinds of polycrystalline silicon electrodes 11a and 11b of the CCD shift register which constitutes the signal charge transfer section 3a; 7 also a contact for connecting together the two kinds of polycrystalline silicon electrodes 11a and 11b of the CCD shift register; 8 an aluminum wiring line to which clock pulses for driving the signal charge reading sections 2a and 2b are applied; 9 a contact for connecting the aluminum wiring line 8 with polycrystalline silicon electrodes 11b for constituting the signal charge reading sections 2a and 2b; and 12 an element separation region for separating each pixel of the light receiving section and the signal charge transfer section.
A size of one pixel of the light receiving section is determined by a portion held between the aluminum wiring lines 5 and 8.
Light shielding of the signal charge transfer section 3a is usually performed by another kind of aluminum wiring line, but its explanation will be omitted.
As apparent from FIG. 12, main factors for deciding a line-to-line distance (distance from the center of the light receiving section 1a to the center of the light receiving section 1b) are as follows:
(1) a size of one pixel of the light receiving section;
(2) a size of the signal charge reading section;
(3) a size of the signal charge transfer section; and
(4) a size of the element separation region between the signal charge transfer section and its adjoining light receiving section array (including a size of the pulse line). In the example of FIG. 12, a line-to-line distance is totally 64 xcexcm (M=8), which is obtained by adding up the following sizes: a size of one pixel of the light receiving section is 8 xcexcm; a size of the signal charge reading section is 10 xcexcm; a size of the signal charge transfer section is 18 xcexcm; a size of the element separation region between the signal charge transfer section and its adjoining light receiving section array is 25 xcexcm; and a size of a connected part between (2) and (3) is 3 xcexcm.
Regarding the foregoing factors (1) to (4), for (1), its size cannot be changed, because it is a fixed pixel size. For (2), it is very difficult to set its size to be lower than 10 xcexcm, because a region must be provided for connecting a clock wiring line for driving the signal charge reading section with a polycrystalline silicon electrode for constituting the signal charge reading section. For (3), easy reduction in size of this portion will cause characteristic deterioration, because as this size is reduced, a maximum signal charge amount processed by the signal charge transfer section is reduced more and a dynamic range for output signals is also reduced more. For (4), it is very difficult to change (reduce) its size, because as in the case of (2), a very large region (about 25 xcexcm) must be provided for connecting the aluminum wiring line 5 as a pulse line on the element separation region 12 with one 11b selected from the two kinds of polycrystalline silicon electrodes of the signal charge transfer section 3a by the contact 6 and also connecting together the two kinds of polycrystalline silicon electrodes 11a and 11b by the contact 7.
In other words, size changes are difficult for any of the factors (1) to (4). In the color linear image sensor shown in FIG. 11, the line-to-line distance having the value shown as an example is substantially a current minimum.
With the color linear image sensor constructed in the foregoing manner, while the signal charges stored in the respective light receiving sections are read by the signal charge reading sections to the specified signal charge transfer sections, some signal charges are left unread (left untransferable). Consequently, a problem of residual images is created. Next, this problem will be described with reference to the accompanying drawings.
FIG. 13 is a sectional view taken along a line Axe2x80x94A of the conventional example of FIG. 12. In the drawing, the same reference numerals as those used in FIG. 12 denote the same elements. Other constituting elements will now be described based on a manufacturing process.
A reference numeral 13 denotes an N-type semiconductor substrate. A P well 14 is first formed on this N-type semiconductor substrate 13 by ion-implanting P-type impurities such as boron and then performing thermal diffusion. Subsequently, an N-type region 15 is formed by ion-implanting N-type impurities such as phosphorus or boron and then performing thermal diffusion. Likewise, an N-type region 17 is formed by ion-implanting N-type impurities such as phosphorus or boron and then performing thermal diffusion. Then, an element separation region 12 is formed by ion-implanting P-type impurities, and reference voltages (0V) are taken in from the outside for the element separation region 12 and the P well 14.
Reference numerals 11a and 11b denote polycrystalline silicon electrodes, which are pattern-formed on the wafer by using a thermal oxidized film 18 as an insulating film after the formation of the element separation region 12. A reference numeral 16 denotes a P-type region for constituting the light receiving section, which is formed, after the formation of the polycrystalline silicon electrode 11b, by performing ion implantation with this polycrystalline silicon electrode 11b as a mask and in self-alignment with the same. Here, incident lights are subjected to photoelectric conversion by a PN junction constituted by the P-type region 16, the N-type region 15 and the P well 14, and thus an operation of the light receiving section is realized. In addition, a buried channel type transistor is composed of the P well 14, the N-type region 17, the thermal oxidized film 18, and the polycrystalline silicon electrode 11a, and thereby the signal charge transfer section 3a is constituted. A reference numeral 20 denotes a flattening film necessary for patterning the aluminum wiring lines 5 and 8.
FIG. 14 shows channel potentials for the light receiving section 1a, the signal charge reading section 2a, the signal charge transfer section 3a, and so on, of FIG. 13. In the drawing, TGoff indicated by a dotted line and TGon indicated by a solid line represent channel potentials for the respective sections at the time of TGoff and TGon of the timing chart shown in FIG. 15.
As apparent from the channel potentials of FIG. 14, signal charges Q are always passed through a portion 2a1 of the signal charge reading section 2a where the N-type region 15 exists when the signal charges Q stored in the light receiving section 1a are transferred (read) therefrom to the signal charge transfer section 3a. A channel potential of this portion is higher than a channel potential of a portion 2a2 of the signal charge reading section where the N-type region does not exist but the P well 14 exists. Consequently, some signal charges are left untransferable (left unread) xcex94Q in the portion 2a1. These signal charges left untransferable xcex94Q will cause generation of residual images during signal charge reading of next time and thereafter, because they are read bit by bit.
For eliminating such signal charges left untransferable xcex94Q, it is possible to employ a structure like that shown in FIG. 16 where the N-type region 15 is separated from the polycrystalline silicon electrode 11b of the signal charge reading section 2a and thereby the N-type region 15 is prevented from existing in the signal charge reading section 2a. Even in the case of this structure, however, since the P-type region 16 is formed in self-alignment with the polycrystalline silicon electrode 11b of the signal charge reading section 2a as described above, a portion 1a1 having the P-type region 16 but not the N-type region 15 is always created in the light receiving section 1a. Consequently, transfer of signal charges from the light receiving section 1a to the signal charge reading section 2a will be impeded. This situation is shown in FIG. 17. FIG. 17 shows channel potentials for the light receiving section 1a, the signal charge reading section 2a and the signal charge transfer section 3a, and so on, of FIG. 16. In the drawing, TGoff indicated by a dotted line and TGon indicated by a solid line represent channel potentials of the respective sections at the time of TGoff and TGon of the timing chart of FIG. 15. As apparent from the channel potentials of FIG. 15, a channel potential is 0V in a portion 1a1 of the light receiving section 1a where the P-type region 16 exists but the N-type region 15 does not exist. Consequently, since the light receiving section is not dependent on a value of a pulse applied to the signal charge reading section 2a, the signal charges Q cannot be read therefrom to the signal charger transfer section.
For the purpose of solving the problems of signal charges left untransferable from the signal charge reading section to the signal charge transfer section and residual images thereby generated, an example realized by using a structure different from that shown in FIG. 13 was disclosed in Patent Publication Hei 8(1996)-10760. FIGS. 18 and 19 are views respectively showing FIGS. 6 and 7 of Patent Publication Hei 8(1996)-10760 referenced herein. But in FIGS. 18 and 19, a number 100 is added to each reference numeral of FIGS. 6 and 7 of Patent Publication Hei 8(1996)-10760. This example of the related art is different from the foregoing conventional example in that a photoelectric conversion section 112 equivalent to the light receiving section 1a of the example shown in FIG. 13 and a charge storage section 113 are separated from each other, and this charge storage section 113 is arranged between the photoelectric conversion section 112 and a transfer gate section 114. A P-type region 104 for constituting the photoelectric conversion section 112 and a P-type region 105 for constituting the transfer gate section 114 are formed in self-alignment respectively with a storage gate electrode 107 and storage gate and transfer gate electrodes 107 and 108. Accordingly, as shown in FIG. 19, in a region from the photoelectric conversion section 112 to below a gate electrode 109-1a of a CCD shift register, no portions exist where a channel potential well (e.g., the portion 2a1 of FIG. 14) or a channel potential barrier (e.g., the portion 1a1 of FIG. 17) is formed. Therefore, no signal charges are left untransferable, and no residual images are generated. In this example of the related art, a negative voltage (e.g., xe2x88x928V) sufficient for setting a potential of a substrate surface below the storage gate electrode 107 equal to a substrate potential (e.g., 0V) is applied to the storage gate electrode 107. This is for the reason that for smoothly transferring signal charges from the photoelectric conversion section 112 to below the gate electrode 109-1a of the CCD shift register, a channel potential "PHgr"ST below the storage gate is set deeper than a channel potential "PHgr"PD of the photoelectric conversion section 112 and shallower than a channel potential "PHgr"TGH below the transfer gate electrode 108 when an H level potential is applied to the same. It is therefore apparent that as long as each of the channel potentials ("PHgr"PD, "PHgr"ST and "PHgr"TGH) satisfies "PHgr"TGH greater than "PHgr"ST greater than "PHgr"PD, a potential applied to the storage gate electrode 107 or the transfer gate electrode 108 can be arbitrarily changed.
However, if the foregoing color linear image sensor is constructed by using the structure of this example of the related art, a line-to-line distance is increased much more compared with that of the conventional color linear image sensor shown in FIGS. 11 and 12. This problem will now be described with reference to the accompanying drawings.
FIG. 20 is an entire constitutional view showing an example of a color linear image sensor constructed by using a linear image sensor of the structure shown in the example of the related art. In the drawing, the same reference numerals as those used in FIG. 11 denote the same elements. In FIG. 20, reference numerals 10a to 10c denote signal charge storage sections to which DC voltages VST are applied. Reference numerals L6 to L11 denote pulse lines for driving the CCD shift registers of the signal charge transfer sections 3a to 3c. (Clock pulse is xcfx861 or xcfx862: connection from the pulse line to each signal charge transfer section is indicated by an arrow).
As can be understood from comparison between FIGS. 20 and 11, a difference in structure between the color linear image sensor of this example and that of the conventional example shown in FIG. 11 is dependent on, in addition to the existence of the signal charge transfer sections respectively adjacent to the light receiving sections of respective colors, the existence of only one or two pulse lines for driving the CCD shift register of the signal transfer section, e.g., 3a, between the color light receiving sections adjacent to each other, e.g., 1a and 1b. For further clarifying this difference, comparison will be made between FIG. 12 and FIG. 21 which is an expanded view of a region surrounded with a broken line X2 of FIG. 20.
In FIG. 21, the same reference numerals as those used in FIGS. 12 and 20 denote the same elements. In FIG. 12, the polycrystalline silicon electrode 11a selected from the two kinds of polycrystalline silicon electrodes 11a and 11b for constituting the signal charge transfer section 3a and adjacent to the polycrystalline silicon electrode 11b of the signal charge transfer section 2a is connected, by the contact 7, with another polycrystalline silicon electrode 11b on a channel stop 12 in a side of the signal charge transfer section 3a where the light receiving section 1a does not exist. This polycrystalline silicon electrode 11b is connected, by the contact 6, with the aluminum wiring line 5 of the pulse line L2 to which a clock pulse xcfx861 is applied. For the remaining polycrystalline silicon electrodes 11a and 11b for constituting the signal charge transfer section 3a, the aluminum wiring line 5 of the pulse line L1 (clock pulse is xcfx862) arranged in a side of the light receiving section 1a where the signal charge transfer section 3a does not exist is connected with the polycrystalline silicon electrode 11a by the contact 6. This polycrystalline silicon electrode 11a itself is passed on the channel stop 12 as a wiring line from the pulse line L1 to the signal charge transfer section 3a and then connected with the polycrystalline silicon electrode 11a of the signal charge transfer section 3a. This polycrystalline electrode 11a is then connected, by the contact 7, with the remaining polycrystalline silicon electrode 11b on the channel stop 12 in the side of the signal charge transfer section 3a where the light receiving section 1a does not exist. It should be noted that the polycrystalline silicon electrode 11a can be directly connected with the polycrystalline silicon electrode 11a of the signal charge transfer section 3a after being passed from the pulse line L1 on the channel stop 12 because the polycrystalline silicon electrode 11b of the signal charge reading section 2a and the polycrystalline silicon electrode 11b intersecting the former on the channel stop 12 are not identical polycrystalline silicon electrodes. For the light receiving sections 1b and 1c of the other two colors, the signal charge reading sections 2b and 2c and the signal charge transfer sections 3b and 3c, one (11b) selected from the two kinds of polycrystalline silicon electrodes can also be passed as a wiring line on the channel stop 12 for separating each pixel of the light receiving section. Accordingly, in this conventional example, it is only necessary to arrange one pulse line between the light receiving sections adjacent to each other, to which clock pulse xcfx862 is applied.
On the other hand, in the case of the color linear image sensor shown in FIG. 21, different from the example of FIG. 12, a signal charge storage section 10a is provided between the light receiving section 1a and the signal charge reading section 2a. For a polycrystalline silicon electrode for constituting this signal charge storage section 10a, one identical to the polycrystalline silicon electrode 11b of the signal charge reading section 2a cannot be used (if an identical electrode is used, short-circuiting or a gap may apparently occur between both). Thus, of the two kinds of polycrystalline silicon electrodes, a remaining one 11a must be used. Therefore, different from the example of FIG. 12, a pulse line to which a clock pulse xcfx862 is applied cannot be arranged in a side of the light receiving section where the signal charge transfer section does not exit and thus the polycrystalline silicon electrode 11a cannot be passed on the channel stop for separating each pixel of the light receiving section 1a so as to connected with the polycrystalline silicon electrode 11a of the signal charge transfer section 3a. For this reason, the pulse line L7 to which a clock pulse xcfx862 is applied must be arranged between the signal charge transfer section 3a and its adjoining color light receiving section 1b as in the case of the pules line L6 to which a clock pulse xcfx861 is applied. As a result, in the color linear image sensor of FIG. 21, since two pulse lines for driving the CCD shift register of the signal charge transfer section 3a must be provided between the light receiving sections adjacent to each other, e.g., 1a and 1b, a line-to-line distance between the light receiving sections adjacent to each other is increased more compared with that of the conventional example shown in FIG. 12. Specifically, a line-to-line distance is totally 96 xcexcm (M=2), which is obtained by adding up the following sizes : (1) a size of one pixel of the light receiving section is 8 xcexcm; (2) a size of the signal charge reading section is 10 xcexcm; (3) a size of the signal charge transfer section is 18 xcexcm; (4) a size of the element separation region between the signal charge transfer section and its adjoining light receiving section array (including sizes of the two pulse lines) is 49 xcexcm ; (5) a size of the signal charge storage section is 8 xcexcm; and (6) a size of a connected part between (2) and (3) is 3 xcexcm. These sizes are obtained based on the numerical values of the factors for calculating the line-to-line distance of the conventional example of FIG. 12. This distance value is an increase of 50% compared with that of the conventional example of FIG. 12. Therefore, with the present example, residual images can be reduced, but it is not suited as a color linear image sensor.
According to an aspect of the present invention, provided is a color linear image sensor having a small line-to-line distance, which comprises on a semiconductor substrate: a plurality of light receiving section array; signal charge storage sections respectively adjacent to the light receiving sections; signal charge reading sections respectively adjacent to the signal charge storage sections; and signal charge transfer section for transferring signal chargers read by the signal charge reading sections from the signal charge storage sections to signal charge conversion sections. In this case, each of the signal charge reading sections includes an N-type (or P-type) region selectively formed in a surface of a P-type (or N-type) semiconductor well, a first P-type (or N-type) region selectively formed in a surface of the N-type (or P-type) region and an electrode formed on the first P-type (or N-type) region interposing a gate insulating film. The electrode of the signal charge reading section is electrically connected with an electrode for constituting a portion of the signal charge transfer section which is adjacent to the signal charge reading section. Driving pulses for both electrodes are made common (xcfx861 (TG)) Not a DC voltage but a pulse is applied to an electrode for constituting the signal charge storage section.
According to another aspect of the present invention, provided is a color linear image sensor having a small line-to-line distance, which comprises on a semiconductor substrate: a plurality of light receiving section array; signal charge storage sections respectively adjacent to the light receiving sections; signal charge reading sections respectively adjacent to the signal charge storage sections; and signal charge transfer sections for transferring signal charges read by the signal charge reading sections from the signal charge storage sections to signal charge conversion sections. In this case, each of the signal charge reading sections includes an N-type (or P-type) region selectively formed in a surface of a P-type (or N-type) semiconductor well, a first P-type (or N-type) region selectively formed in a surface of the N-type (or P-type) region and an electrode formed on the first P-type (or N-type) region interposing a gate insulating film. An electrode for constituting the signal charge storage section is electrically connected with an electrode for constituting a portion of the signal charge transfer section which is not adjacent to the signal charge reading section. Not DC voltages but common driving pulses xcfx862 (ST) are applied to both electrodes.
With the color linear image sensors constructed in both of the foregoing manners, residual images can be reduced without greatly increasing a line-to-line distance between the light receiving section array adjacent to each other.